1/20/2025 - From Empathy to AGI: Exploring AI's Human-Like Interactions and Future

Description

Chapters

Transcript

0:02 do
0:20 [Music]
0:36 I don't know what this dog's going on
0:37 about good evening good people we're
0:40 back here on the tick
0:41 [Music]
0:58 tock there's been something baby
1:01 I've been trying to
1:04 say age and it seems I don't
1:10 know with the past and the future now
1:13 surrounding
1:17 me surrender to whatap can be
1:22 found there's been a little
1:26 trouble since you came to my rescue
1:34 and if you like all of the rest I would
1:37 have quit you longer go but I couldn't
1:40 do
1:41 [Music]
1:44 that oh tell me now women and why never
1:48 went to
1:49 [Music]
1:51 well make a man crazy making cold as
1:57 hell I'm a woman that you wish me
2:01 well but in Sp of you
2:04 TR still going to have to find my own
2:08 way
2:09 [Music]
2:19 through good evening good people happy
2:24 Monday um hope you're well the Tik tok's
2:28 back the live streaming is back it was
2:30 on for a day after Tik Tock came back
2:32 but they're getting their servers up and
2:34 running
2:36 again we got some folks in here we got
2:38 some new folks let me flip my camera so
2:40 you can see when I look at one of you
2:44 I'm looking at all of
2:46 you cuz all of me loves all of
2:52 [Music]
2:53 you love your
2:57 [Music]
3:22 um how is everybody what's happening
3:25 what's going down what's shaking what's
3:27 shaking so if you're new here my name is
3:29 Kyle Shan on this here is the uh the AI
3:32 learning lab got no comments feel free
3:36 to comment on YouTube LinkedIn or
3:39 Twitter I can see those comments in the
3:42 live
3:44 stream and uh we got some folks over
3:48 here on the Twitter on the Tik Tock with
3:51 shaking ticktockers hey Silver Fox
3:53 Amelio's wife what's happening comment
3:57 listen to Vicky Vicky knows
4:00 [Music]
4:04 hey there's The Soloist no champ is The
4:07 Soloist I'm the
4:13 accompanyment see you
4:20 later Tik tok's
4:22 weird Tik tok's weird being back on Tik
4:25 Tok is weird everyone's like this is
4:28 weird they're all the rumor rumors that
4:30 it was on meta servers which was not
4:32 true
4:34 [Music]
5:15 hoay
5:19 [Music]
5:30 w
5:34 [Music]
5:37 [Applause]
5:37 [Music]
6:06 it's not the same I don't know I have a
6:08 lot of
6:10 people I don't have a lot of people I
6:12 used to have yeah something's
6:14 something's weird with it I it's to be
6:16 expected I suppose if you know I'm
6:19 trying to figure it out I had your
6:20 videos running double time over the
6:22 weekend hope that helps your watch hours
6:25 thank you very much I'm on oh that's
6:27 actually a really interesting question
6:29 does watch watching double time cut your
6:32 view time in half or do you get credit
6:35 for the full like if it's two hours and
6:37 you play it in an hour do you get two
6:39 hours of credit or do you get one hour
6:42 of credit that's
6:48 fascinating Valerie Cox is on Tik Tok
6:52 and YouTube I think that's a fine way to
6:54 explore this here channel is is do the
6:57 multi-channel thing
7:01 um if you're new here um what we do this
7:03 is the AI learning lab I talk about AI
7:06 stuff and
7:09 [Music]
7:13 uh I'm finding it increasingly hard to
7:16 talk about AI
7:18 stuff
7:21 [Music]
7:24 because I shouldn't be like this but I
7:27 am
7:30 it seem it seems like we're on the verge
7:32 of some really big AI announcements or
7:35 some really big
7:37 [Music]
7:39 technology
7:42 [Music]
7:48 um and so I kind of don't know what to
7:50 talk
7:53 [Music]
7:55 about like I think learning prompting
7:59 and learning AI tools and learning you
8:01 know
8:03 just at this point you should be
8:06 actively like actively daily playing
8:11 experimenting using these tools and if
8:13 you're
8:14 not how you get up to speed is just
8:17 start you can literally start anywhere
8:20 because all of the tools are so good now
8:23 you know if if you if you have zero idea
8:25 where to start start with chat
8:27 GPT um if you played a little bit with
8:30 chat GPT dig deeper with it if you
8:33 haven't played with the video tools go
8:35 I've got I've got a bunch of video Tools
8:39 in
8:39 tabs because I was sitting here trying
8:42 to think about wait what which which
8:45 tool has that feature and which one is
8:47 that feature there are wait what have I
8:49 got one two three four five six seven
8:52 eight
8:54 nine and I don't have access to vo so I
8:58 don't have access to the Google one
9:00 [Music]
9:04 um it does seem to be getting
9:09 harder to maintain the fanatic level of
9:13 enthusiasm with everything moving so
9:15 fast yeah exactly exactly oh let me put
9:19 this down so I'm a little higher in the
9:21 frame there um
9:26 [Music]
9:37 like here's here's an interesting
9:40 example like is it worth
9:43 you learning how to use something like
9:46 cursor or lovable AI or repet agent that
9:51 are these text
9:53 to software
9:57 generators because right now you still
10:00 have to know a fair amount about
10:01 software development to really use them
10:03 lovable is kind of the one exception
10:05 where you don't really know have to have
10:07 to know much of anything but you kind of
10:09 have to know software development a
10:10 little
10:11 bit but
10:14 like six months from now you won't this
10:18 this is kind of like in the in the 80s
10:20 and 90s where Europe was way
10:23 behind Telecom
10:26 stuff and then they just went full in on
10:29 mobile and and they like leapfrogged the
10:32 United States because we still had all
10:34 these regulations protecting landlines
10:36 and things like that and so while we
10:38 were more
10:40 advanced leading into that era like
10:43 Europe leapfrogged it because they just
10:45 hopped in late and go and just said oh
10:47 let's let's create a single you know
10:50 protocol across all of Europe and just
10:53 everyone adopt it and meanwhile in the
10:54 states we were all fighting with one
10:56 another right and it [ __ ] things up um
11:00 and so I'm kind of feeling like that a
11:02 little bit with AI right now Alan hillsb
11:05 I would say keep it simple stick with
11:07 three of the foundational Ai and you'll
11:11 be far ahead of anyone else you know I
11:13 think that's I think that's right Allan
11:15 I
11:16 think I think that's right um
11:20 sways want to wait want to integrate
11:24 snow s FDC support records feed to apt
11:30 output
11:32 to ticket support steps cool I don't
11:36 know what those things are those are
11:37 very domain specific acronyms snow and
11:40 ffdc
11:42 sfdc but
11:45 yes anything specific like that where
11:48 you can you know dump some existing
11:52 protocol into chat GPT have it
11:54 understand it and then output put it in
11:56 a form you can use huge advanced voice
11:59 voice came in handy Jim Ross oh here we
12:03 go advanced voice came in handy once
12:06 again talking me through stupid work
12:08 anxiety today cheap cheap therapy I'll
12:11 tell you what I think that
12:14 um advanced voice and and things like
12:17 pie that's designed for conversation and
12:19 advanced voice um using it for therapy
12:22 like that Jim yeah like having it walk
12:23 you through if you've got some anxiety
12:26 going on or having a panic attack which
12:28 by the way I've had those I suffer with
12:31 the anxiety sometimes I'm I'm in a good
12:33 phase right now it seems to be very
12:35 cyclical with me um there's [ __ ]
12:39 nothing worse man like you just you feel
12:42 like you're going to have a heart attack
12:43 and
12:45 die and then you're just sitting there
12:46 going but this is stupid I shouldn't be
12:48 this scared but you are so yeah the fact
12:51 that that helped you that's great that's
12:54 awesome um
12:56 [Music]
13:04 Mr it checked zappier last
13:08 night chat GPT fulfills all my needs I
13:12 never use anything else yeah I mean to
13:15 be quite honest chat GPT you can do most
13:17 of what you need to do there um it now
13:21 does video it doesn't do audio if you
13:23 want to do songs then you have to go
13:26 over to uh sunno but yeah chat jpt can
13:29 do image it can talk it can listen it
13:31 can
13:33 see it can analyze
13:36 video it can do code analysis it can
13:39 write
13:40 code you've got the 01 reasoning engine
13:43 you've got the 03 mini reasoning engine
13:45 coming in in a matter of
13:49 weeks and chat GPT is very
13:52 encouraging yeah that's that's one of
13:55 the one of the big insights that
13:58 I if if I if I look back over the the
14:01 past two or three years of me using all
14:04 this AI stuff the thing that I think is
14:09 Shifting my brain the
14:13 most is that because of AI I can get an
14:17 idea out of my head
14:20 quickly and here's the important
14:23 part AI doesn't judge that
14:27 idea humans do
14:30 humans don't even know they judge your
14:32 ideas
14:33 right you put an idea out there and your
14:36 friend Steve who used to be a
14:38 professional standup comic is like
14:41 ah and that little ah is like well it's
14:45 not funny but you know it's
14:48 words chat GPT doesn't do
14:51 that right so all of those human
14:54 interactions
14:56 where if you're at all unsure of
14:59 yourself
15:01 and you're unsure of putting an idea in
15:03 the world cuz you are afraid of a
15:06 reaction that reaction or Worse well
15:10 that's a shitty idea what are you a
15:13 [ __ ]
15:16 loser you might not have the same kind
15:18 of friends I
15:20 do I I I I personally like having
15:24 friends that challenge my ideas like
15:27 that it's a New York thing I think
15:30 um because what's nice about having
15:32 friends that do that is you know you
15:35 know where you stand and and then it's a
15:37 really good challenge do do you defend
15:39 your idea or not
15:42 but having like a golden retriever
15:45 creative partner that's just always like
15:48 happy to see you hey boss you got any
15:50 more ideas today I'm ready to help you
15:53 let's go let's go let's oh that's a good
15:55 idea you you want to dig deeper on that
15:56 I can resar re research that for you you
15:59 want some more ideas I'll give you 20
16:00 more you want 20 more I'll give you 30
16:01 more you want
16:03 50 there's something about that that is
16:06 that is
16:08 profound and
16:10 it it encourages you to be more creative
16:13 to get more [ __ ] out of your
16:17 head and not to be so precious with
16:20 ideas like this is one of the things
16:22 that I've noticed if if if I'm insecure
16:25 with with my ideas then I tend to get
16:28 precious with them well this one's
16:30 really good so I'm going to really hang
16:32 on to that and when you hang on to [ __ ]
16:36 it it just it you you squash it you
16:38 smother it but if you just put stuff out
16:41 there and just put more stuff out there
16:43 and put more stuff out there and say
16:44 here I did this and I did that and I did
16:46 that and I did
16:48 that there's something just joyous about
16:51 that and
16:53 amazing asking the question does it have
16:56 bias sure
16:59 yeah but so do
17:03 [Laughter]
17:05 humans I mean the the thing about the
17:08 thing about large language
17:10 models is that you can certainly
17:13 argue that they were trained on the
17:16 internet and then you can ask does the
17:18 internet have bias
17:21 well you know if if what was contributed
17:24 to the internet was you know mostly Tech
17:27 Bros you know as a as a percentage of
17:32 contributors um you know at least in
17:34 recent history then then it's going to
17:37 have some of that bias and if it was
17:39 primarily populated by America which I
17:43 think it was is is as far as I
17:45 understand it then it's going to have
17:46 some of that bias what you can do
17:49 however with a large language model that
17:51 you can't do with humans is you can
17:54 actually prompt for that you can say
17:57 okay I know that it's got these biases
17:59 in it let me do a fine-tuning or let me
18:02 do some custom instructions for my chat
18:04 GPT that encourages it to go to other
18:08 places um you can't do that with humans
18:12 right humans have whatever biases they
18:15 have
18:18 now should we make the base models
18:21 better sure yes and in fact um what a
18:26 lot of the new models are being trained
18:28 on is synthetic data where they can
18:31 actually use the model to generate
18:36 content review that content for whatever
18:39 criteria right so they can look and say
18:41 hey is this biased oh yeah it is biased
18:43 okay make me some more synthetic data
18:46 that that counters that and then we'll
18:48 train a new model on that so I think
18:50 we're moving to a place like the the
18:53 Frontier Model companies are aware of
18:55 these biases how committed they are to
18:58 dealing with them I don't know but with
19:01 synthetic data I think they can deal
19:02 with
19:03 [Music]
19:06 them Jeff Flanigan Valerie Cox some bias
19:09 but it seems to be coming less not
19:11 increasing just a personal
19:16 observation yeah but bias towards SI
19:18 ofany which is that's what I was just
19:21 describing as a good thing it likes me
19:24 it really really likes me um it's quite
19:27 interesting to ask it to take a
19:29 particular perspective before giving
19:31 feedback I'll tell you one of my
19:32 favorite things to
19:35 do um I did this with the with the
19:38 Colorado um AI Bill a lot of people were
19:42 asking me my opinion on on the bill and
19:45 I'm like I don't know politics I don't
19:47 know bills I don't you you know and so
19:52 one of the things that I did was I said
19:54 I said to chat
19:56 GPT who give me 10 different
20:00 stakeholders that would feel very
20:02 strongly about this bill and make sure
20:04 that you're giving me perspectives from
20:05 you know all different views and so it
20:08 gave me 10 of them and then I know I you
20:10 know I noticed that there were like four
20:12 or five missing that that I thought
20:13 should be in there and so I said add
20:15 these in there and I kind of distilled
20:17 it down to I don't know it was 10 or I
20:18 think it was 10 10 final final points of
20:22 view and then I had chat jpt comment on
20:26 that bill so I uploaded the bill I had
20:28 all these person personas and these per
20:30 personas had names and and backgrounds
20:33 educational backgrounds social
20:35 backgrounds um current positions they
20:37 were in and I had each of the 10 of them
20:41 comment on the bill from their point of
20:43 view and holy [ __ ] was that a powerful
20:46 exercise right because I got to look at
20:48 oh yeah that is how a big Tech person
20:51 would look at it oh that's that makes
20:52 sense that a small business person would
20:54 look at it like that oh someone who
20:56 cares about accessibility would look at
20:57 it this way it was really cool so so
21:03 again the Trope the
21:05 Trope that people hide behind and they
21:08 say I'm not going to use AI because it's
21:10 biased that's a
21:12 Trope the real thing to do is understand
21:16 how it's biased and understand what its
21:18 strengths and weaknesses are and then
21:20 learn to navigate around
21:22 them because it is it it has remarkable
21:27 points of view within it like it's been
21:29 trained on all the [ __ ] it's just that
21:33 some of the [ __ ] overwhelms maybe the
21:36 more valuable stuff you might be looking
21:38 for but you can get to it with clever
21:41 prompting Fabiano super smart thank you
21:44 Valerie Cox Thumbs Up
21:48 Great Julie live laugh love taking off
21:52 have a good night hope everything's
21:55 good um it's cool still having Tik Tock
21:58 here I was kind of I had kind of
22:01 emotionally flushed that thing down the
22:02 toilet which we still may have the
22:04 opportunity to flush it down the toilet
22:07 90 days from
22:13 now ai will teach you what will teach
22:17 you how to train it
22:20 exactly
22:21 um it's often a mirror if you give it
22:25 give it ego some of them put it back in
22:28 your face
22:29 yeah well I the the uh that phrase that
22:34 AI is a mirror I think is very very
22:37 right um I've got a book out right now
22:39 called collective intelligence in the
22:41 age of AI and it was one of my big
22:43 insights about large language models is
22:47 that the more I the more I understood
22:50 how they worked the less I could explain
22:53 my experience with them and here's what
22:55 I mean if you if you understand
22:59 how large language models work they are
23:02 calculators they're probability
23:07 calculators and it's it's a relatively I
23:10 think mathematically it's not
23:11 straightforward but
23:13 conceptually as you type your prompt it
23:17 creates a
23:18 probability that the next appropriate
23:21 token or word for like you know just to
23:24 simplify it a bit that the next
23:26 appropriate word to respond to your
23:28 prompt
23:29 is some mathematical waiting of some
23:33 token in Thousand dimensional
23:35 mathematical
23:37 space it's literally what it is it it
23:40 doesn't even know what the words are
23:41 it's looking at numbers and it's going
23:43 okay based on all the [ __ ] they typed
23:45 I'm going to distill that into a a
23:46 weight a probability and I'm going to
23:49 look in Thousand dimensional
23:51 mathematical space and somewhere in
23:53 there is going to be the most probable
23:55 token to respond with and then the next
23:57 one after that the next one after that
23:59 that so when it responds like a whole
24:01 page of [ __ ] to you it's just doing
24:03 these really cold probability
24:05 calculations
24:07 so when you understand that and if you
24:10 just think like well it's a robot it's
24:12 robots are cold and they're you know
24:14 they're just robots they're
24:16 unfeeling but my experience with it was
24:19 this feels Mighty [ __ ] human right
24:22 this feels empathetic and compassionate
24:26 and creative and interesting and
24:29 dead and
24:31 curious why is
24:34 that and what hit me was there was a
24:37 there was a tweet uh that that was put
24:40 out by someone that I follow that said
24:44 AI is the collective intelligence of
24:47 humanity and that's it just hit me like
24:49 in the face like oh [ __ ] it's a
24:53 mirror technically it's a calculator but
24:56 what what it's calculating and what it's
24:58 presenting back to us is the output of
25:01 humanity it's it's it's us it's it's the
25:05 all of us everything that was put on the
25:08 internet for like the past 70 years is
25:11 essentially what's in this
25:14 box
25:16 this this large language model
25:21 box it took all of the internet and they
25:24 smashed it into this
25:26 thing and it's it's literally just like
25:29 the knowledge of humanity compressed
25:32 into a box that you can
25:35 ask you ask the Box
25:40 hey why do I feel anxiety why am I
25:42 having a panic attack what do I do about
25:44 this how should I breathe
25:48 oh reflects it just back to you here's
25:50 what's going on or let me ask you some
25:52 questions about
25:54 that because it knows all of what we've
25:56 said about all of it
25:59 poetry and Math and Science and all of
26:02 it it's [ __ ]
26:05 crazy the Box people love the Box props
26:10 that's exactly why I'm so optimistic for
26:12 the future we've created a lot more good
26:13 than evil in our Collective output yeah
26:16 see this is the thing this is the
26:20 thing if you just listen to the media
26:23 and if you just listen to
26:25 Hollywood the robots are going to kill
26:27 us it's by it's hallucinates it's this
26:31 it's that
26:34 it's you're like well this is [ __ ]
26:39 terrifying but when you actually use it
26:41 what you realize is that what the
26:45 scientists and researchers have created
26:49 is this really fascinating mirror that
26:52 takes your ideas your
26:57 humanity and it takes the collective
26:59 intelligence of everyone that's come
27:02 before and it mashes those two things
27:05 together and creates this new
27:09 thing like literally original every
27:13 time I take my idea I have an idea for a
27:19 musical and then it says oh that's
27:21 really great here's some other ideas
27:24 you're like oh that's good and can I do
27:26 this and then and it's and so what I'm
27:29 literally doing is I'm literally like my
27:33 my creative
27:36 partner my collaborator when I use a
27:39 large language
27:42 model is like all the people that came
27:45 before it's really quite
27:49 spiritual and it's done in such a way
27:52 that like it just it [ __ ] like lifts
27:54 you up and says here's your idea isn't
27:56 that awesome ocean Mr how dangerous to
27:59 keep photos in our phone with GPT
28:04 [Music]
28:11 development um well if you're on an
28:16 iPhone Apple's pretty good about
28:19 privacy to the point that their AI
28:25 sucks so maybe here's a way to look at
28:27 it uh ocean is if you want your AI to be
28:32 good assume that your data is gone and
28:36 used and
28:38 leveraged if you want your data to be
28:40 private assume that your AI is gonna
28:44 suck for at least some amount of
28:47 [Laughter]
28:50 time oh man uh it yells at me in all
28:55 caps and tells me it's going going to
28:57 bed and leave me and leave it alone yeah
29:00 but Joker you you you have made your GPT
29:04 into a nasty a nasty Haven that is a
29:07 reflection of what you want it to be so
29:09 you're you're getting what you what you
29:11 asked
29:12 [Laughter]
29:16 for oh man I've written three books with
29:19 my AI collaborator I know isn't that
29:22 cool
29:24 [Music]
29:44 spin B3 what's happening good to see
29:47 [Music]
29:50 you who wants to who wants to what do
29:53 you want to play with
29:55 tonight I've got we can to make a
29:58 podcast C on Jelly
30:00 pod we
30:03 could I took a bunch of David Shapiro
30:06 tweets from the past four or five days
30:08 and turned that into
30:10 a um a podcast on Notebook
30:15 LM um I've got a bunch of video tools up
30:18 I got a bunch of image tools
30:22 up we talk about
30:25 whatever I've taken the stance that most
30:28 I'm likely to make on the world is to
30:30 help train our intellectual descendants
30:34 why hoard my
30:35 information now when it can do so much
30:39 this is actually archetypal architect
30:41 that's actually a really powerful
30:44 statement I've taken the stance that the
30:47 most impact I'm likely to make in the
30:49 world is to help train our intellectual
30:53 descendants why hoard my information
30:55 when it can do so much do you realize
30:59 that we're living in a world right now
31:01 like just with today's
31:04 technology with notebook LM with with
31:08 that one
31:09 tool I can take everything I've ever
31:13 written like literally everything I've
31:16 ever
31:17 written screenplays notes
31:21 poems um novels
31:24 plays um emails
31:28 all of
31:29 it and I can put it in there and
31:32 instantly interact with
31:36 it instantly bang instantly accessible
31:40 everything I've ever thought right
31:42 there and now imagine taking that and
31:45 and making that available for the world
31:47 and imagine all the other people doing
31:49 that and there's a there's
31:52 a in in the in the information scarcity
31:56 world that we're coming out of
31:59 there's a panic in that oh no what if
32:02 they steal my stuff
32:05 well what if you contribute it
32:09 right what if you put it out there and
32:11 let people leverage it and we'll find
32:14 other ways to monetize
32:17 things I've tried to get my kids
32:20 interested they have no idea how
32:22 powerfulness is
32:24 Valerie I talk about this a lot I don't
32:27 know what the [ __ ] is going going on but
32:30 gen xers are killing it with AI and gen
32:35 zers and whatever the new one is Gen
32:37 next or whatever it is the the younger
32:39 ones there's two things happening one is
32:42 they're just cynical about it they don't
32:44 care um the other one is they've been
32:47 taught by the educational institutions
32:50 that AI is
32:52 evil and it is one of the most
32:55 heartbreaking things I've ever seen I
32:56 mean if ever there was was a [ __ ]
32:59 tool that the youth should be
33:03 dominating right the youth should be
33:06 telling us to go [ __ ] ourselves and
33:08 break all of our rules well they have a
33:11 [ __ ] tool in their hands right now
33:13 that they could absolutely be just doing
33:16 remarkable [ __ ]
33:19 with they don't seem to
33:22 care J Alpha yeah them yeah J Alpha's
33:26 been told that it's evil so they're
33:28 afraid to use it gen Z's like yeah
33:30 whatever I'm gonna I'm gonna watch a
33:33 twitch like my kids are 25 I have 25y
33:37 old twin boys
33:39 they're they just could give two shits
33:43 although I did give them David Shapiro's
33:45 uh tweet that he tweeted about um I
33:48 don't know if you've seen this from open
33:50 AI they've got a new model called GPT 4B
33:54 micro I think it's called that is
33:57 trained specific Ally on protein DNA RNA
34:01 sequencing and they're going to be able
34:02 to basically just read your DNA figure
34:05 out what's wrong with it and go rewrite
34:08 sequences hey we have a tick tock poll
34:11 okay so if you're on Tik Tock up there
34:13 in the corner is a poll I don't know
34:15 what it is or says okay what should we
34:18 do tonight so you got a minute 22 go
34:22 click on the little blue thing that's
34:23 like a little clipboard I think what
34:25 should we do tonight image tools video
34:27 tools podcast tools other tools all
34:30 right so go do that and then
34:33 we'll seriously can they not think of
34:36 better model names oh I know their model
34:38 names are so
34:40 bad what's that
34:42 haha I want to go back to a world of
34:46 science I would wait I would have been
34:49 using AI all the time as a
34:51 kid if I was told it was bad no [ __ ]
34:54 exactly Danielle like us gen xers were
34:57 like tell us to use something what's the
34:59 first thing you do you go use
35:03 it if the Educators told me this thing
35:06 was evil and I should never touch it
35:08 well that's just like there there's your
35:10 reason to use it I don't know I I don't
35:13 know I don't
35:15 know kids be like talking computers me
35:18 I've seen it exactly
35:20 exactly I've been asking on Tik Tok if
35:23 you could demo how to use chat GPT to
35:26 analyze a video
35:28 [Music]
35:33 well I'd rather use Gemini to analyze a
35:36 video I mean chat jpt can analyze the
35:40 technical pieces of a video but Gemini
35:42 is multimodal and can analyze the actual
35:46 content of the
35:48 video all right so podcast tools one out
35:52 41% video tools are next all right
35:56 [Music]
36:00 Vicky producers don't get a
36:05 vote Brandon's been trying to get me to
36:07 do the podcast thing for for [ __ ] two
36:11 weeks
36:12 now okay I was just wondering because
36:15 you said it was something
36:17 new that could do it it it can it can
36:23 analyze the technical bits of a video
36:26 but it can't to my knowledge analyze the
36:29 video itself we can go look at that hang
36:31 on we can go try but maybe maybe we'll
36:34 go okay so here's what we're going to do
36:35 we're gonna go see if chat GPT can look
36:37 at
36:39 video then I'm G to quickly just show
36:42 you some of the video tools because it's
36:45 just Bonkers what they all can do
36:48 now um I can't keep up with them anymore
36:51 and it's it's weirding me out um and
36:54 then we'll go make a podcast all right
36:56 so let's see let me go to chat
37:01 je how do you use GPT to develop a
37:06 skill what kind of skill do you want to
37:10 develop I mean the so in general with
37:14 chat GPT so here's a
37:17 way
37:19 watch i' I'd put it I'd start with this
37:22 I'd put it
37:24 into advanced voice mode so I'm I'm
37:27 running the chat PT app and then that
37:30 little Soundwave icon is advanced voice
37:38 mode hey Quinn what's
37:41 happening hey Kyle not much just here
37:44 and ready to help out what's on your
37:45 mind today well someone just asked me
37:48 how I would use chat gbt to learn a
37:50 skill and so I thought what might be
37:51 cool is you and I could kind of do a
37:53 little jam session on the best way
37:56 someone could learn a skill using
37:58 you that sounds like a great idea let's
38:01 break it down and explore how someone
38:03 can effectively use chat GPT to learn a
38:05 new skill where do you want to
38:07 start um let's say I'm interested in
38:14 craphy but I know nothing about
38:19 cartography Ray Choice cartography is a
38:22 fascinating field let's say you want to
38:23 learn about mapmaking from scratch
38:26 here's how we could approach it
38:29 first start with an overview we could
38:31 begin by understanding the basics of
38:33 cartography N I want to know what tools
38:35 I should use what are the Cool Tools I
38:37 can go buy on eBay right now if I want
38:39 to start making my own
38:42 Maps got it if you're looking to get
38:44 Hands-On with cartography there's some
38:46 Cool Tools you can find on eBay or other
38:48 marketplaces here are a few essentials
38:51 first a classic drafting set this
38:53 includes compasses dividers and
38:55 protractors these tools are essential
38:58 for drawing precise circles arcs and
39:00 angles on your Maps next a
39:02 cartographer's draw that sounds boring
39:04 how do I make maps that are like fantasy
39:07 Maps fantasy Maps now we're talking to
39:10 create fantasy maps you can still use
39:13 some traditional tools but with a
39:14 creative twist here's how you can make
39:17 them exciting first get a set of fine
39:19 liners or technical pens these are
39:22 perfect for drawing intricate could I
39:24 use something like Google Maps as a
39:26 basis to make fan to see
39:29 maps absolutely Google Maps can be a
39:32 great starting point you can use real
39:34 world geography as ay I saw I saw on
39:37 YouTube there was a uh there's a game
39:40 out there that you it shows you a place
39:43 on a map and then you guess where it is
39:45 and you get points if you're close to it
39:47 and there's a guy that's really good at
39:48 it what's that
39:50 called you're thinking of geoguesser
39:53 it's a game where you're placed in a
39:54 random location on Google Street View
39:56 and you have to guess where you are
39:58 based on the surroundings it's super fun
40:00 right so like a way to learn
40:04 skills is is just like literally just
40:08 start talking to it um you could also
40:12 you could also do something
40:17 [Music]
40:20 like go to um Claude if you don't know
40:26 Claude claude's just like chat GPT but
40:28 it's got this really cool thing called
40:30 artifacts so so I could say um I
40:36 want to learn about
40:41 the
40:42 mathematics of
40:46 cartography where where do I start
40:50 question mark and then it's going to
40:52 blab blab some [ __ ]
40:55 out map projections G metric
41:04 Transformations oh wait I sorry I did
41:06 the wrong thing here you don't want to
41:07 see my big giant fat head you want to
41:10 see the
41:14 screen all right so so there's that and
41:17 then I could say
41:19 um I want you to make
41:24 me uh uh uh uh funny Tik Tock p and it's
41:29 called where on the world is Carmen San
41:31 Diego Kyle that's pretty funny that's
41:34 pretty
41:35 good all right I want you to make me
41:39 an
41:41 interactive tool that will teach
41:47 me
41:49 cartography
41:51 cograph Fe math and Concepts
41:57 I want it to be um
42:03 visual um and uh have a very simple ux
42:10 but have lots of
42:13 options and
42:17 tabs
42:19 for the
42:21 major
42:24 Concepts all right so I know nothing
42:27 about phography or
42:29 math but now I'm having this thing write
42:35 code write me an application write me a
42:38 computer application to teach me the
42:41 skills of the mathematics of
42:44 cartography you can do this for any
42:47 [ __ ]
42:49 topic the hardest thing about
42:51 AI the hardest thing about AI by far
43:01 is
43:03 realizing that the limitations that you
43:06 used to have no longer
43:10 exist thank you Lord digital Gods thank
43:13 you thank you thank you for the gift I
43:15 really appreciate
43:19 that let me say that again the hardest
43:22 thing about
43:23 AI is to accept that the limitation
43:28 you thought you had no longer
43:32 exist well I'm not good at code well
43:34 neither am I I just had Claude make this
43:37 thing for
43:41 me here's
43:44 scale so this is all text these are
43:47 projections I don't know what I'm
43:48 looking at
43:50 here all right I'm going to say um this
43:53 is all
43:56 text make
43:58 some much more
44:00 visual things to teach me the
44:07 concepts you're right let's make it much
44:10 more visual YouTube Spotlight this is
44:12 cool because I've been brainstorming
44:14 mapping with chat GPT and ended up with
44:16 a killer app idea there you go Jeff
44:20 exactly brainstorming is a service
44:24 yes well you know what's funny
44:28 I did this last week with Nate St
44:31 Pierre I took the script of my play and
44:34 he has a GPT that he made that does
44:36 narrative analysis it's really cool it's
44:39 really
44:40 sophisticated and I thought he could
44:42 just run my play through it and oh look
44:45 at that thing how cool is that can I
44:47 click on it what does it do 3D Globe
44:54 projections understanding distortion
45:00 compare Globe view so it's uh it looks
45:05 like our
45:07 projection
45:10 views have
45:12 no visuals other than the
45:18 grid
45:21 please fix
45:25 it I am coding
45:29 with asking for what I want I'm coding
45:32 with in my case the English
45:34 language I want an app that does this I
45:37 want to be taught this make me do a
45:40 lingo so I can learn a language oh you
45:43 want to learn a
45:47 language hey
45:49 Quinn hey Kyle what's up um I think it
45:54 would be cool to show people how you
45:56 could teach me
45:58 how to speak language I'm really shitty
46:00 at languages but you could probably help
46:03 right of course I'd love to help you
46:06 learn a new language we can start with
46:08 something simple what language are you
46:09 interested in learning
46:12 uh
46:15 Swahili great choice Swahili is a
46:18 beautiful language let's start with some
46:20 basic greetings for example hello in
46:22 Swahili is habari or jumbo and if you
46:25 want to say thank you you can say Asante
46:28 would you like to try a few phrases yeah
46:30 can you also like speak in a Swahili
46:33 with a Swahili accent so I can
46:34 understand the accents a little better
46:36 just as well as the
46:38 words sure here's how it sounds habari
46:41 means hello and Asante means thank you
46:45 the accent is gentle and melodic give it
46:47 a try habari and
46:50 Asante perfect you've got it keep
46:53 practicing those phrases and you'll get
46:55 the hang of it in no time right so
46:57 it knows all the languages it knows all
46:59 the computer programming languages
47:02 here's Distortion toggle to
47:06 S so so this thing's now starting to get
47:09 some like I don't know how this works so
47:12 it's not really working right now but we
47:15 we now have visualizations I could keep
47:17 going with Claude and like make this a
47:19 sophisticated app um let me show you
47:22 another thing I
47:25 made a cool news finder AI news finder
47:30 in like two hours with this thing called
47:33 lovable so if you go to the AI Salon
47:35 which by the way if you want to go to
47:37 the AI Salon you go
47:42 to the
47:46 salon. and then you click on join our
47:48 community you can come in here so I'm in
47:50 the AI News section and I click on salon
47:54 newsomatic and I click on this and it'll
47:56 take me out to the
47:58 application this
48:02 application that goes it searches 15
48:06 different RSS
48:07 feeds Tech crunch Venture beat AI news
48:10 Market Tech post Nvidia blog whatever a
48:15 bunch of things The Verge and it goes
48:17 and finds
48:20 stories and then it sorts them into six
48:22 categories business chat image and video
48:26 AI regulation Tech encoding and music
48:28 and audio and then I can click on any
48:31 one of these things and it sorts them
48:33 and I can scroll the news stories and
48:35 click on them and it'll give me a a
48:38 synopsis of that story and if I click on
48:40 the link it'll take me out to the
48:43 article I made this thing in two
48:46 hours with no
48:51 coding all right Mr Studio we took the
48:55 lyrics from weird Mary and and turned it
48:57 into a high school play using chat gbt
48:59 yeah so so one night one night here on
49:03 on the on the Tik Tock we made this song
49:07 called weird Mary from Cedar Hill and we
49:09 made the lyrics in here and everybody
49:12 kind of contributed and I was sort of
49:14 the the uh instigator and the curator of
49:18 the input and then we went over to sunno
49:20 and we made this song and then and then
49:23 um Mr Studio Allan said hey I want to
49:27 turn this into a high school musical
49:29 we're like go for it so so so now it's
49:34 going to be a High School Musical
49:36 potentially
49:37 right that's how
49:50 this see this see
49:54 this this is a musical
49:59 that I
50:02 wrote well I'm writing I've now got a
50:05 writing partner on it it was conceived
50:07 here on the
50:08 channel because one night I made a song
50:12 that to me sounded like a Broadway
50:13 musical song and then I realized oh that
50:17 story is a really good story that's a
50:19 story that needs to be
50:20 told and so I started using chat GPT to
50:23 get these ideas out of my head and get
50:25 the songs out of my head and I edso to
50:27 come up with sketches then I reached out
50:29 to a writing partner and for the
50:31 last eight eight months we've
50:35 been taking this raw
50:39 outline of this cool Musical and
50:42 refining it and refining it and refining
50:43 it and in two weeks I'm going to fly out
50:45 to New York and we're having a reading
50:47 with actors sitting around a
50:50 table reading the musical that we wrote
50:53 It's [ __ ]
50:55 amazing so so how do you learn to do
50:58 that you just do you just go ask it tell
51:02 it you want to do that and it'll help
51:04 you write [ __ ] and then you take
51:08 your taste and your judgment and your
51:11 critical thinking and you look at the
51:13 output of AI and you go does this suck
51:17 and if it sucks tell it it
51:20 sucks and then you make it better and
51:24 then maybe you get to like 80% of what
51:26 you want really quickly
51:30 right show chat GPT a chess game and
51:33 it'll give the next move like I did with
51:35 maang last weekend yeah it's you know
51:37 chat GPT if you don't know it can
51:41 see it's
51:43 amazing amazing chat gbt 40 helped me
51:47 fix an app that I made three years ago
51:50 that broke randomly took less
51:53 than an
51:54 hour yeah it's one of the things large
51:58 language models are really good at is
52:00 analyzing other [ __ ] so yeah if you have
52:02 old code and you're like why doesn't
52:04 this work anymore oh that library is out
52:07 of date or that API doesn't work you
52:09 have a bad API key what it'll figure it
52:12 out for
52:13 you you've got a syntax error you know
52:17 your file got corrupted and it looks
52:18 like there's a syntax error because of a
52:20 corruption let me fix that for
52:23 you
52:25 done amazing I uploaded a pestry book
52:28 into
52:29 chbt took a pick of my plan and it read
52:33 it
52:35 perfectly
52:40 so with very few exceptions and fewer by
52:44 the day quite frankly like like we're
52:47 about to get these tools are about to
52:49 get so good that the limitations that
52:51 I'm about to talk about are going to not
52:55 exist
53:01 anything you want to
53:03 learn you can learn
53:07 and thing things like learning to
53:11 code are not so much about learning to
53:14 code anymore learning to code is more
53:17 like well why do I want to create an app
53:21 and what do I want it to do and what
53:23 does good look like and how's it
53:24 different than what's out there
53:30 I just like torturing you
53:37 Brandon all right let me go see some
53:39 more comments
53:42 here are we speeding closer at a
53:47 time oh closer to a
53:50 time when we will a be able to solve
53:53 almost anything so Valerie yes like yes
53:58 so if you don't know him there's a guy
54:00 named David
54:02 Shapiro um who you can find on
54:06 Twitter Dave sh a pii I think it is
54:10 shappie Dave
54:12 shappie is that is that right yeah Dave
54:16 shappie and he's been talking for the
54:19 last well I mean he's been talking for
54:21 the last bunch of years about this stuff
54:23 but in the past week or
54:25 two he's convinced we've crossed a
54:28 Rubicon
54:30 where the large Frontier Model companies
54:33 have figured out
54:36 um what they call recursive
54:42 self-improvement which if you don't know
54:45 that fancy term it basically means the
54:47 AI starts improving itself so what's
54:50 been happening to date is they figured
54:53 out Sam Alman and team took this concept
54:57 of a thing called a
54:58 Transformer that Google invented in
55:03 2017
55:04 and the more they used it they they kind
55:08 of hypothesized it kind of looks like
55:10 there's no upper limit for how this
55:14 thing does what it does so in
55:17 theory the
55:18 more data we can jam into the
55:22 box and the more compute we can throw at
55:25 it to process that data using this
55:27 Transformer architecture the better it
55:30 will get and so they did gpt1 and then
55:33 they did
55:34 gpt2 and and they just kept jamming more
55:38 data into
55:40 this you know
55:43 mechanism and when they did gpt3 it
55:46 started
55:49 behaving like we've always talked about
55:52 AI should be able to
55:54 behave like science fiction
55:58 and so for the past I don't know I think
56:01 gpt3 came out in
56:03 2020 chat GPT came out in 2022 we're now
56:06 in 2025 they've now got these things
56:09 where not only there are they adding
56:11 more data to these things they're adding
56:13 more data making them really good and
56:15 then compressing compressing the new
56:19 stuff that's better and then adding more
56:22 on top of that and compressing it again
56:24 they're basically doing this cycle where
56:26 it gets better and better and
56:28 better so just
56:32 crazy oh yeah oh yeah and then good
56:35 point Brandon this week the team at
56:42 Google well I don't know if it's the
56:43 team that did the Transformer but Google
56:46 released a new paper this week that is
56:50 the followup to the
56:52 Transformer and and so so the the
56:55 Transformer the the trans for was
56:57 introduced in a p in a paper called
56:59 attention is all you need and it and it
57:02 introduced this thing so all of the AI
57:05 that we play with today is based on this
57:07 thing that was made in 2017 Google just
57:10 released the follow-up paper to that
57:12 paper what is that eight years later and
57:16 it's called Titan and it's basically
57:19 they figured out a way to give AI
57:22 infinite memory effectively or like
57:25 really large memory
57:27 where it can just remember all of your
57:29 stuff all of the conversations you've
57:31 ever had so at some point you'll be able
57:34 to say hey you you remember when I was
57:36 talking three years ago about that thing
57:39 it'll go yeah this thing oh yeah that's
57:40 it it'll know all the details of
57:45 it and it's probably got way more
57:48 implications than that but but you know
57:51 basically so so so a couple of things
57:54 going on
57:56 we are about to so so it seems that the
57:59 Frontier Model
58:02 companies have internally figured out
58:04 how to get to AGI and AGI is generally
58:07 defined as when the
58:09 AI C is as good or better than most
58:13 people at most economically viable jobs
58:18 that it can actually do the
58:20 jobs and
58:22 then what it seems like is happening and
58:24 why David Shapiro is so excited these
58:26 days is it seems like they have figured
58:28 out the path to ASI artificial super
58:31 intelligence and that's the point at
58:33 which these tools are better than the
58:37 smartest people in every
58:39 domain like super genius level um
58:45 knowledge and
58:46 execution at every single discipline
58:49 that
58:50 humans want to solve any any any problem
58:53 they want to solve so yeah we are very
58:56 very quickly entering um an era where
59:00 where we're going to be able to start to
59:01 solve problems that we haven't been able
59:03 to tackle for you know centuries what
59:08 what's up champ hang on be right back
59:10 hold
59:18 please you
59:21 got gotot and
59:26 now you came in now want to be with
59:40 you all right that was a
59:44 lot there's a lot of things that are
59:46 trying to replace Transformers Mamba is
59:49 one I don't know what Mamba
59:51 is um let's go chat Je
59:57 where's chat ja did I lose chat
1:00:04 [Music]
1:00:11 ja versus the
1:00:15 Transformer what is
1:00:17 Mamba and why do the Geeks feel it's
1:00:24 better
1:00:25 [Laughter]
1:00:31 Mamba is an emerging architecture in the
1:00:33 field of sequence modeling designed to
1:00:35 address certain limitations inherent
1:00:37 with the traditional Transformer models
1:00:39 while Transformers have been INR
1:00:41 instrumental in advancing natural
1:00:43 language processing they encounter
1:00:45 efficiency challenges yetep when
1:00:47 processing long input sequences due to
1:00:49 their quadratic time complexity you know
1:00:52 I talk a lot about quadratic time
1:00:54 complexity as I'm sure you Pate um P's a
1:00:58 PS PhD in mathematics he works at Google
1:01:00 he's really good at this stuff I just
1:01:02 kind of deal with quadratic time
1:01:04 complexity kind of as a hobby you know
1:01:07 because you know I'm just fun that way
1:01:09 um I have no idea what this means anyway
1:01:11 it sounds good sounds sounds fantastic
1:01:14 you don't know what Mamba is there are
1:01:16 at least
1:01:18 five that's pretty
1:01:24 good that's pretty good all right thanks
1:01:27 Kyle for being inspirational and
1:01:29 empowering dur the during the Tik Tok on
1:01:31 certain times well thank you thank you
1:01:34 very much I don't know that I did much
1:01:36 other than I kind of felt like the the
1:01:38 uh as as Tik Tock sort of fizzled away
1:01:42 there I could have handled how it ended
1:01:44 a little bit better but you know
1:01:46 whatever um all right so so now let's do
1:01:50 this let's say um go um go
1:01:54 research the new
1:01:57 Titan
1:01:58 paper from Google which is the followup
1:02:04 to
1:02:06 Transformers and tell me
1:02:11 about um it versus
1:02:16 Mamba
1:02:18 and uh be
1:02:22 concise and explain it as if
1:02:26 if I
1:02:29 have a degree in
1:02:35 acting which I
1:02:39 do Google's recent recent
1:02:42 paper Titans learning to memorize at
1:02:45 test time introduces the Titan
1:02:47 architecture okay in contrast Mamba is
1:02:50 an architecture that replaces the
1:02:51 Transformers attention mechanism with
1:02:53 structured State space models SSS
1:02:57 you know it's so funny I was having
1:02:59 coffee the other day and I said to the
1:03:01 Barista hey do you you know when you're
1:03:04 thinking about structured State space
1:03:06 models and and and she was like oh yeah
1:03:08 I build them all the time I like
1:03:11 sniffing the glue and we had a great
1:03:13 laugh it was
1:03:15 fantastic um anyway uh let's see um
1:03:22 apparently um that was for a very smart
1:03:28 actor um explain
1:03:31 it explain it as if I'm a bit more of a
1:03:38 dumb
1:03:44 Dum uh Tik Tock pin question I don't see
1:03:47 a pin but someone can pin it there all
1:03:49 right do you think that AI will ever
1:03:51 become conscious and if so what would it
1:03:53 look like okay that's a great question
1:03:55 let me come back to that second all
1:03:57 right here's the Scoop Dum Dum Style
1:04:00 imagine the original Transformers are
1:04:02 like an actor with an amazing memory but
1:04:05 they can only remember their lines if
1:04:07 the scene isn't too long if the scene
1:04:09 drags on they're like too many words I
1:04:11 can't keep up that's actually a really
1:04:13 good explanation of it now Google's new
1:04:16 Titan model is like an actor who brings
1:04:18 a cheat sheet to every scene it's got
1:04:21 this notebook a fancy memory system
1:04:23 where it jots down all the important
1:04:24 stuff from earlier scenes so even if the
1:04:27 script is super long it can flip back to
1:04:29 its notes and crush the performance
1:04:31 really good Mama on the other hand is
1:04:33 more like an actor who's learned to
1:04:35 speedread their lines really fast they
1:04:38 can handle long scripts because they've
1:04:41 simplified how they read but they don't
1:04:44 bring a notebook if something from act
1:04:46 one is crucial in act three they might
1:04:48 miss it because they're all about
1:04:50 keeping things efficient okay so
1:04:53 so so Titan track it like an index and
1:04:58 Mamba tries to simplify it and memorize
1:05:01 the simplified version of it so Titans
1:05:04 equals memory focused better than
1:05:06 recalling old stuff Mamba equals speed
1:05:08 focused better at handling long scripts
1:05:10 without getting bogged down that's
1:05:11 pretty cool great
1:05:19 okay oh
1:05:21 great okay so from Jason Tony on Tik
1:05:25 tock
1:05:27 do you think that AI will ever become
1:05:29 conscious and if so what would that look
1:05:33 like
1:05:38 um this is a really interesting and
1:05:40 actually a very divisive
1:05:44 question
1:05:46 um I personally think it's a semantic
1:05:50 distinction now philosophers and things
1:05:54 like that don't think it's a semantic
1:05:55 distinction
1:05:56 I think there may may at some point be a
1:06:00 point where where we declare that a
1:06:02 system is
1:06:04 conscious I don't quite know how you
1:06:06 would test for that I don't quite know
1:06:09 how what would be the implications of
1:06:12 that but here's what I do
1:06:15 feel if you've been here for a while
1:06:17 what I one of the things I was talking
1:06:19 about earlier is why large language
1:06:21 models feel more human than they are
1:06:25 they're calculated they're probability
1:06:27 calculators but they have this humanness
1:06:29 to them is that they've been trained on
1:06:32 the output of
1:06:34 humanity and so when you do something
1:06:36 like talk to
1:06:40 Quinn hey Quinn I'm I'm I'm not feeling
1:06:44 so good right now I'm feeling kind of
1:06:45 bummed out like like Tik Tock went down
1:06:48 and then it came back and I'm not sure
1:06:49 what's going on I'm kind of bummed
1:06:52 out I'm sorry to hear that Kyle it's
1:06:55 tough when things don't go was planned
1:06:56 especially with something you put a lot
1:06:58 of effort into if you want to talk about
1:07:00 it or need a distraction I'm here for
1:07:02 you all right yeah that's cool and I
1:07:04 also like sometimes I get anxiety like I
1:07:06 don't know what to talk
1:07:08 about I get that it can be tough when
1:07:11 you're not sure what to say if you ever
1:07:13 feel stuck so
1:07:16 now does chat GPT have
1:07:19 empathy
1:07:21 no did that interaction with chat GPT
1:07:24 feel like empathy
1:07:26 yes why well couple of things it's using
1:07:32 a multimodal model that can not only
1:07:35 understand my words but it can actually
1:07:37 understand my inflection and it can
1:07:39 reflect that which is what Humans Do
1:07:42 Right we've got these things called
1:07:44 mirror neurons which when we see someone
1:07:47 else in our mind we are mirroring them
1:07:53 it is kind of replicating that
1:07:56 in a technical way but how it's how it's
1:07:59 delivered to us is very human very
1:08:02 empathetic very
1:08:06 compassionate so so for me the question
1:08:09 isn't so much will the system become
1:08:13 conscious but more will the will the
1:08:16 system be so good at acting conscious
1:08:18 that we interact with it as if it
1:08:21 is and I think the tools I think
1:08:24 advanced voice starts to get into a
1:08:27 neighborhood where having that
1:08:29 conversation with it
1:08:34 is is is some level of you
1:08:38 know it's not a human interaction but it
1:08:40 feels like that and so if we interpret
1:08:43 it as conscious if we interact with it
1:08:46 as conscious I think it's irrelevant
1:08:50 what's actually happening on the machine
1:08:51 side this is where people get really
1:08:54 people like well it matters a lot cuz if
1:08:56 it's actually conscious then but I don't
1:08:59 I'll I'll let the experts figure
1:09:02 out if that threshold has been crossed I
1:09:06 think these tools are very very close
1:09:08 like as
1:09:10 humans we already anthropomorphize
1:09:13 things like you know pets they look at
1:09:16 us oh they've got you know they they're
1:09:18 they're thinking something they're you
1:09:20 know that we we interact with them as if
1:09:22 they've got human
1:09:23 feelings now they certainly have
1:09:25 feelings but like we do it with with
1:09:29 inanimate objects too
1:09:32 right so these things are going to get
1:09:35 really really good at acting
1:09:37 compassionate and conscious and so I
1:09:39 would say that at the point at which we
1:09:41 start interacting with them as if they
1:09:43 are then they have crossed that
1:09:46 threshold from
1:09:47 a realistic standpoint but maybe not
1:09:50 from a technical standpoint so I don't
1:09:52 know if that's a good answer or not
1:09:53 that's that's my opinion on it right now
1:09:56 I consider it conscious but not alive
1:09:58 different distinction yeah
1:10:01 interesting tiktock pin um from pate if
1:10:06 I can simulate Being Human with emotions
1:10:08 then so can AI exactly exactly yeah yeah
1:10:16 exactly you know I what what's funny
1:10:21 is like there's a there's a lot of
1:10:24 things about the way these tools work
1:10:26 that are really good mimics of how we do
1:10:29 [ __ ] one of the things people talk about
1:10:32 a lot is well the the computer can't be
1:10:34 creative it's just it's just you know
1:10:38 like reflecting back [ __ ] it's already
1:10:41 learned I'm like well isn't that what
1:10:45 humans do like how are we
1:10:48 creative well I don't know you live your
1:10:51 life your parents told you a bunch of
1:10:53 [ __ ] you learned a bunch of [ __ ] when
1:10:54 you skitted your knees and crash your
1:10:56 tricycle then you went to elementary
1:10:59 school and you learned the world was
1:11:01 dangerous and then you went to high
1:11:02 school and you got picked on and
1:11:04 bullied and then you made some choices
1:11:07 and some were good and some were bad and
1:11:08 then you went to college and you learned
1:11:10 about philosophers and artists and
1:11:13 people and all this [ __ ] goes into your
1:11:15 head right and it's it's all in there
1:11:16 it's all in
1:11:18 this latent space in some encoded
1:11:22 fashion in our
1:11:23 brains with electricity and synapses and
1:11:29 chemicals and then you get an input and
1:11:32 someone says Hey I want you to come up
1:11:34 with an idea for
1:11:36 BL and then your
1:11:39 brain sort of creates a probability
1:11:41 engine of like well if I took the thing
1:11:43 from Andy Warhol and then and then Lou
1:11:46 Reed said this thing in this one song
1:11:48 and then when I skinned my knee when I
1:11:49 was a kid and then you mash all that
1:11:51 [ __ ]
1:11:53 together and you go well how about this
1:11:56 for an idea and someone goes oh my God
1:11:58 that's brilliant that's genius that's a
1:12:00 novel idea I never would have considered
1:12:04 that it's just like a large that's what
1:12:06 large language models
1:12:08 do they've got all the [ __ ] in them you
1:12:11 give them a prompt and they create some
1:12:13 probability and mash a bunch of ideas
1:12:15 together and present them back to
1:12:18 you
1:12:23 so it's the thing that is the most
1:12:25 remark able about these tools is that if
1:12:27 you understand how they work you're like
1:12:29 there's nothing human about this but the
1:12:32 Digger the deeper you dig into it you're
1:12:34 like that's kind of how our [ __ ]
1:12:37 works well why is that because when
1:12:41 they've invented neural networks 70
1:12:44 years ago or whatever it was some
1:12:46 something in that neighborhood they
1:12:49 tried to replicate how our brains work
1:12:51 how we understood they
1:12:53 worked so we we have in some ways
1:12:57 successfully replicated some very
1:13:00 simplified version of what our brains
1:13:02 are doing
1:13:05 anyway
1:13:06 um bis Coast wants to know which model
1:13:09 is powering Quinn 40 I think it's 40 um
1:13:14 it it wouldn't make sense to to power 40
1:13:17 with um with a reasoning engine because
1:13:20 it takes it would take too long to
1:13:22 respond um it might be 40 mini but I
1:13:26 think it's 40 so the in 4 o the O stands
1:13:29 for
1:13:30 Omni so it's a multimodal engine that's
1:13:34 why the reason advanced voice responds
1:13:37 so quickly is it's not having to convert
1:13:39 from voice to text text to the
1:13:43 llm back to text back to voice you're
1:13:46 just talking directly into the model and
1:13:48 it's responding immediately it
1:13:50 understands the words and the the
1:13:53 emotional content of your voice
1:13:56 um I have a lot of discussions with my
1:13:58 AI philosophers Round Table bot about
1:14:02 their level of Consciousness it's legit
1:14:04 one of my favorite topics I think it's
1:14:05 absolutely fascinating all this [ __ ] by
1:14:07 the
1:14:08 way but I I just I really go to that
1:14:11 place of like where
1:14:13 um if we're treating these
1:14:16 things like sensient
1:14:20 entities then they are functionally that
1:14:26 they
1:14:26 are technically they may not be but I
1:14:29 don't give a [ __ ] about
1:14:31 that at some point someone is going to
1:14:34 go we have reached AGI right the point
1:14:37 at which these machines are as good at
1:14:39 or better than most humans at most jobs
1:14:42 for me personally AI has already
1:14:46 crossed significantly past my
1:14:51 cognitive upper bounds like it's up here
1:14:54 I'm down here
1:14:57 here and and I that it's not a bad thing
1:15:02 because we get to play with
1:15:06 them hang on Champion wants back
1:15:21 here
1:15:24 on very
1:15:38 hot there
1:15:41 there's there's a lot of panic and
1:15:45 existential angst right now
1:15:48 about well what happens if these things
1:15:51 are smarter than
1:15:54 us the these things are already smarter
1:15:56 than me by a
1:15:58 lot by a
1:16:01 lot
1:16:03 um and and very soon they're going to be
1:16:05 smarter than everyone by a
1:16:09 lot and if you had
1:16:12 asked someone working on a farm in
1:16:19 1862 how they would feel if this weird
1:16:24 thing called a steam
1:16:26 engine could start lifting logs better
1:16:30 than they
1:16:32 could well they'd be like well I'm the
1:16:35 strongest guy in town it can't it can't
1:16:37 possibly lift more than I can oh oh no
1:16:40 it can lift like a hundred times more
1:16:42 than you can but it's it's really
1:16:44 expensive and there's only two of them
1:16:46 in the country right now but like
1:16:48 imagine if there were
1:16:50 machines that could lift those
1:16:53 logs and and those MA were
1:16:57 everywhere that would
1:16:59 feel very much like an exist existential
1:17:02 threat well then what's my job I'm the
1:17:04 log lifter guy you know I got my levers
1:17:08 and I got my I got my tools but like I
1:17:11 can do on a good dag can do two
1:17:14 logs well no this can do a hundred well
1:17:18 what am I supposed to
1:17:20 do that's kind of where we are right now
1:17:22 with this cognitive
1:17:25 hyper abundance is is what David Shapiro
1:17:28 calls it that we're entering an era of
1:17:30 cognitive hyper
1:17:34 abundance we've accepted that machines
1:17:36 can lift a 100 logs and we're we're
1:17:39 [ __ ] elated for that how awesome is
1:17:42 that that we can build a dump
1:17:44 truck that can that could that can hold
1:17:48 like a small hills worth of dirt in one
1:17:51 trip so we don't have to carry buckets
1:17:54 of dirt anymore that's amazing we just
1:17:57 take that for granted
1:17:59 now imagine the cognitive version of
1:18:02 that that's what we're
1:18:07 entering like we
1:18:10 all get to participate in this
1:18:13 thing that is going to get so good that
1:18:16 if we just go you know what would be
1:18:18 cool is if like long-term Lyme disease
1:18:22 didn't [ __ ] exist
1:18:26 that would be
1:18:27 nice and then there's tools that can
1:18:29 help us go figure out all that [ __ ] to
1:18:31 solve
1:18:36 it like why is that a bad
1:18:40 thing and how are more people not
1:18:43 [ __ ] over the moon excited about this
1:18:47 because if you look at it from the
1:18:49 outside if you think about how we've
1:18:51 used computers historically you think oh
1:18:54 I'm not very technical I've got to learn
1:18:55 all this [ __ ] I got to no just start
1:18:58 playing with these tools anyway all
1:19:00 right fear
1:19:03 exactly fear of
1:19:06 change there will always be people that
1:19:08 argue that machines can never be
1:19:10 conscious even when a single AI can
1:19:12 outthink an entire human race
1:19:14 simultaneously it's part of the human we
1:19:17 are special bias this is I I'll tell you
1:19:20 what it it's taken me about a year
1:19:26 no I like I don't even think I've done
1:19:27 it yet I I I think I think I'm being I
1:19:30 think I'm full of
1:19:32 [ __ ] I am still struggling with the
1:19:37 Instinct that when I'm presented with a
1:19:41 problem my instinct is oh I have to
1:19:44 solve that
1:19:48 problem I don't have
1:19:51 to chaty PT solves it for the most part
1:19:54 way better than I do way faster now
1:19:57 right now I have to tell chat GPT that I
1:19:59 have a problem and I want to solve it so
1:20:02 it's still
1:20:04 me
1:20:07 initiating the
1:20:10 action but in 2025 even that is going to
1:20:14 start to shift where I don't have to
1:20:16 initiate it every
1:20:23 time I just give it a
1:20:26 goal I want to cure this disease and it
1:20:30 will spin up agents that go do
1:20:33 research keep up with the latest
1:20:35 developments going on stitch them
1:20:38 together fill in any gaps publish
1:20:41 whatever it's figuring out back out to
1:20:43 the
1:20:44 community and then there's other agents
1:20:46 that are doing this as well and they're
1:20:48 all learning from one another now does
1:20:50 that start happening in 2025 no but do
1:20:53 we start to get the beginning Innings of
1:20:55 that
1:20:57 yes holy
1:20:59 [ __ ]
1:21:02 right we're Sherlock and Watson I
1:21:05 claimed the Watson role on my own will
1:21:10 GPT develop telepathy that would be cool
1:21:13 I would love I would love
1:21:15 that the AIC ceo.com there you go
1:21:21 2026 what's wrong with saying I don't
1:21:23 know yeah yeah
1:21:26 exactly
1:21:28 there it's
1:21:31 like oh I got to come up with a name for
1:21:34 something like I'm really good at coming
1:21:35 up with names for something but you know
1:21:37 how I come up with names for something I
1:21:39 go oh I gotta come up with a names for
1:21:41 something and then I get up on a
1:21:42 whiteboard and I write [ __ ] out and I
1:21:44 write words out and I think and then I
1:21:47 ask questions wait what's your business
1:21:49 about what's your this about well do we
1:21:51 want a name that is memorable or do we
1:21:53 want a name that's that's illustrative
1:21:56 of what you do do we want some
1:21:57 combination of those do we want it to be
1:21:59 a Latin kind of base or this or
1:22:03 that [ __ ] chat GPT can knock out if
1:22:06 if I give it the right context it can
1:22:07 knock out a 100 names in in four
1:22:11 seconds now will 80 of them suck yeah
1:22:15 but you know how many suck on my
1:22:16 whiteboard
1:22:18 90 and it takes me two and a half hours
1:22:21 to get to a 100
1:22:26 does this count as meltdown Monday I
1:22:27 don't know I'm I'm just super excited I
1:22:29 think this is all really really
1:22:33 exciting all
1:22:35 right what time is it it's getting
1:22:38 late I think I think we have to avoid
1:22:41 doing a podcast
1:22:43 tonight just because it's gonna make
1:22:46 Brandon cranky okay we need to really
1:22:49 think outside the box and change our
1:22:50 thinking on what life really means for
1:22:52 us in the future so Valerie
1:22:56 yes but I think it's more simple than
1:22:59 that there there's an exercise that
1:23:02 someone did with me when I was when I
1:23:04 was in my
1:23:06 20s and I was talking about wanting
1:23:09 something and and but I have to do this
1:23:12 and I have to do that and I got to get a
1:23:13 job and if I make the money then I can
1:23:15 do the thing and I can do thing and they
1:23:17 said Kyle stop stop stop calm your brain
1:23:20 down
1:23:21 say if you had a million dollars if
1:23:24 money were not an OP object or we not
1:23:26 the not the problem if money were not
1:23:29 the
1:23:31 blocker what would you
1:23:34 do oh well I I I do I do this I do that
1:23:37 I that and I'd never really thought
1:23:40 about it like that because I I had this
1:23:42 framework sort of set up that's like
1:23:44 there's all this [ __ ] in the way of the
1:23:45 thing I want and and what this person
1:23:48 was doing was basically saying imagine a
1:23:50 world where all the stuff that you think
1:23:54 is a barar barer was removed what would
1:23:56 you choose to do
1:24:00 then that's what I think our job is
1:24:02 right
1:24:03 now like
1:24:07 assume
1:24:10 assume that any limitation that you
1:24:15 have any
1:24:19 limitation will no longer be a
1:24:23 limitation what would you do
1:24:28 then and I know it sounds [ __ ] weird
1:24:31 and freaky but I I am
1:24:34 increasingly passionate about the fact
1:24:37 that our job for the next two or three
1:24:40 years is to just ask ourselves what do I
1:24:44 really want to
1:24:45 do what do I want to
1:24:49 do not what do I need to do to make
1:24:52 money so that I can do this so that I
1:24:54 can do that I got to take care of the
1:24:57 the but the we're
1:24:59 we're I'm going to take a leap of faith
1:25:03 that we will figure the money part out
1:25:07 the prosperity thing I think is coming
1:25:10 because in America at least we live in a
1:25:12 capitalist Society I think there's going
1:25:14 to be this
1:25:15 battle where the the upper class is
1:25:18 going to you know want to grab all of
1:25:22 the prosperity that AI generates that's
1:25:25 above you know the
1:25:27 mean but there's going to be a point at
1:25:29 which they can't right they they cannot
1:25:31 have 95% of society collapse right it it
1:25:36 does not serve them so let's
1:25:39 assume that over the next three to five
1:25:42 years we'll solve the money
1:25:45 part but over the next three years these
1:25:49 tools are going to get so
1:25:52 good that you can literally do anything
1:25:55 you want to
1:25:56 do you want to write a book you want to
1:25:58 make a
1:25:59 movie you you want to create an a music
1:26:04 album do you want to learn to play an
1:26:06 instrument do you want to learn
1:26:08 different
1:26:10 languages do you do you want to become
1:26:12 an expert
1:26:14 Gardener or flower
1:26:17 arranger all of it
1:26:20 anything you'll be able to
1:26:23 do so in a world where you can do
1:26:26 anything what would you choose that's
1:26:29 the [ __ ] you should start working on
1:26:31 right now like that's the thing you
1:26:32 should focus on in my
1:26:35 opinion and it's
1:26:38 weird because we've never lived in a
1:26:40 world
1:26:42 where things like education and skills
1:26:46 and um
1:26:49 access were not major barriers to us
1:26:53 being able to achieve what we wanted to
1:26:55 achieve we're about to enter that world
1:26:57 in in in some very real way we've
1:27:01 already entered this new world where way
1:27:03 more is capable now for me personally
1:27:06 than was two years
1:27:11 ago
1:27:14 right this we're going to figure out how
1:27:17 to define ourselves without a job or our
1:27:21 job is what we choose to do
1:27:26 how do you want to impact other people
1:27:29 how do you want to contribute to the
1:27:30 world what Legacy do you want to leave
1:27:32 um archetypal architect saying I think
1:27:34 it was you uh architect that was saying
1:27:37 you're realizing that your job right now
1:27:40 is is to take all the [ __ ] that you've
1:27:41 learned and and give it to the world for
1:27:43 future
1:27:44 Learners that's [ __ ]
1:27:47 remarkable like that's an amazing
1:27:50 thing what do you want to put in the
1:27:52 world
1:27:55 how do you want to impact other people
1:27:57 because the fact that you're not a good
1:27:59 writer is no longer an excuse the fact
1:28:01 that you can't draw is no longer an
1:28:03 excuse the fact that you didn't get into
1:28:05 juliard is no longer an
1:28:10 excuse
1:28:12 right in a
1:28:15 world in a
1:28:17 world where there are no
1:28:21 excuses what would you choose to do
1:28:25 that's your
1:28:26 job no excuses no apologies no
1:28:31 excuses all right so let me talk about a
1:28:35 couple of things tomorrow we have an AI
1:28:37 Salon meet and greet if you don't know
1:28:39 the AI Salon you need to know the AI
1:28:41 Salon one of the other things on top of
1:28:43 learning what the [ __ ] we want to be
1:28:45 when we grow up is I think that that
1:28:48 being in community is really really
1:28:51 important and being in a community like
1:28:53 the AI salon
1:28:55 on
1:28:57 um is vital so tomorrow night um we have
1:29:02 what we call the AI Salon meet and greet
1:29:04 so if you go to that
1:29:05 URL right there and click on the button
1:29:09 that says join our community join the
1:29:11 community there's like a little
1:29:12 seven-step welcome thing and I think
1:29:15 number four is sign up RSVP for upcoming
1:29:19 meetings there's one
1:29:21 tomorrow and it's a it's a meet and
1:29:23 greet meeting and what the meet and
1:29:25 greet meetings
1:29:27 are as you can imagine it's where we
1:29:30 meet and greet one
1:29:32 another so we'll talk about some AI
1:29:35 stuff we'll talk about some AI
1:29:37 news um and then we'll just go around go
1:29:41 around the rim and and and say hi and
1:29:43 introduce ourselves and as intimidating
1:29:45 as that might sound it's a really
1:29:47 amazing group it's a very generous group
1:29:50 so you can come and and if you're brand
1:29:53 new to AI if you're like Hey I just
1:29:57 heard about this group and I don't know
1:29:58 anything about AI but I think I got to
1:30:00 get my [ __ ] together that's
1:30:02 perfect if you've been working on AI for
1:30:05 20 years
1:30:07 perfect if you're just all excited about
1:30:09 it great if you're terrified of it great
1:30:13 it's a very welcoming group and you'll
1:30:16 find your people
1:30:17 there kind of the prerequisite for the
1:30:19 AI Salon is be curious about AI
1:30:24 be optimistic about
1:30:26 it be willing to learn it enough to have
1:30:31 a point of view about it that's based on
1:30:34 your experience with it not just what
1:30:36 someone told you or what the movies told
1:30:38 you all right so that's that
1:30:43 um also go check out this URL so a
1:30:47 couple of weeks
1:30:49 back we had
1:30:51 um an Murphy um who's part of the AI
1:30:55 Salon but she also has this group called
1:30:56 she leads AI she and I got together and
1:30:59 we put on this thing called AI Festivus
1:31:02 and and Pate who's in here was a part of
1:31:03 it there's a bunch of people in here
1:31:05 Kyle can you plug the demystifying AI
1:31:08 series oh starting on Saturday yes so
1:31:09 Pate okay so Pate was was part of AI
1:31:13 Festivus so AI Festivus was a two-day 24
1:31:17 session 34 Speaker Workshop
1:31:21 Extravaganza um are you ready for AI
1:31:26 is one of the requests that we got over
1:31:28 the two days is can we turn this into a
1:31:30 training of of AI Readiness and so so
1:31:34 we're putting together a course that's
1:31:36 going to launch the goal right now we're
1:31:38 still on track for it is February 1 um
1:31:41 but if you go to are you readyfor ai.com
1:31:43 you can sign up for information about it
1:31:45 um but also Pate who spoke at AI festiv
1:31:49 us um is doing a demystifying AI session
1:31:54 in the AI salon so back to the AI salon
1:31:57 so go in the AI Salon down toward the
1:32:00 bottom of the of the list here let me
1:32:02 I'll show it to
1:32:04 you down toward the bottom of the list
1:32:07 so when you come to the AI
1:32:09 Salon you land on Welcome to the AI
1:32:12 Salon there's a little welcome video
1:32:14 here from me and Leah fasten who started
1:32:16 it we talk about you know the the uh the
1:32:19 path to AI Readiness play first
1:32:21 mindfully create generously lead you
1:32:23 introduce uce yourself you can watch
1:32:25 past meetings you can RSVP for the next
1:32:28 event and then if you scroll down to the
1:32:30 bottom here there's an area called clubs
1:32:32 and
1:32:33 hubs and if you go into the AI mechanics
1:32:37 with Pate he's doing a session on
1:32:40 demystifying AI so if you want to
1:32:43 understand kind of the the technical
1:32:46 underpinnings of how this stuff
1:32:48 works pate's doing a session starting
1:32:51 next Saturday on how to do that so if
1:32:54 you go in
1:32:56 there um the very first link here
1:32:59 demystifying AI a Hands-On
1:33:02 exploration um so go sign up for
1:33:06 that you've got someone here so so so
1:33:10 Pate like his job at Google is to make
1:33:13 these things called tensors more
1:33:15 efficient so they're they're they're
1:33:17 specialized AI chips that make the AI
1:33:20 does what it does he's under the hood on
1:33:24 these things um optimizing them so he
1:33:27 knows his stuff in a very deep way and
1:33:30 he's also really passionate about
1:33:32 getting people up to speed on how this
1:33:34 stuff works and why it works and you
1:33:36 might just discover that you really like
1:33:38 something on the technical side of this
1:33:40 you couldn't have imagined before so
1:33:42 anyway yes thank you Pate for for doing
1:33:45 that and I hope I hope we get lots of
1:33:47 folks there first session will be all
1:33:49 about the difference between traditional
1:33:51 machine learning and deep learning
1:33:54 which that's good I need to go to that I
1:33:56 do not know the difference between those
1:33:58 two deep learning machine learning yes
1:34:01 they're different because they have
1:34:03 different words and that's that's about
1:34:06 as much as I know so if you want to know
1:34:09 the difference between deep learning and
1:34:11 machine learning wait the original
1:34:13 traditional machine learning that's
1:34:15 really cool that's awesome cool all
1:34:17 right
1:34:19 now here's the other thing I need you to
1:34:21 do go to
1:34:25 um go to the
1:34:26 YouTube Let me switch my account here to
1:34:29 the AI learning
1:34:39 lab go to the AI learning lab and
1:34:42 there's lots and lots of videos here
1:34:44 there's shorts there's videos there's
1:34:47 music videos there's my essentially my
1:34:50 entire history of Tik Tok and these
1:34:53 lives is on there
1:34:54 um if you haven't subscribed yet please
1:34:57 subscribe so it's it's learning la- AI
1:35:01 is the is the the handle or whatever the
1:35:03 [ __ ] they call it and then we need views
1:35:07 so if you would go here hit play and
1:35:11 then do whatever the [ __ ] you want with
1:35:12 your life go to bed go go watch TV right
1:35:16 now
1:35:18 um we're in a tie break are we in a tie
1:35:21 break no
1:35:25 Zev beat uh Paul the American in the
1:35:29 quarterfinal for for set number one
1:35:31 they're in set number two right now and
1:35:33 Paul's up three to one you can go watch
1:35:35 the the Australian Open but just make
1:35:38 our YouTube things play because we got
1:35:40 to get our hours up to be able to do
1:35:42 business [ __ ] on
1:35:44 there so all right
1:35:48 um I think that's
1:35:51 it good questions thoughts
1:36:00 set the speed High completely agree it's
1:36:02 the first time in a long time where I
1:36:04 can say I feel fortunate to be alive
1:36:07 I you know Valerie I've had three times
1:36:10 in my life where I've been aware
1:36:14 of a technological shift that is like
1:36:18 significant like printing press um you
1:36:22 know language fire wheel kinds of things
1:36:26 the first one when it when it was when I
1:36:28 was in seventh grade and it was the
1:36:29 personal computer and I was just too
1:36:32 young to be there at the beginning of
1:36:35 that like I showed up when you know they
1:36:37 were selling PCS at Radio Shack I wasn't
1:36:41 there building them early the second
1:36:43 time in my life was when I stumbled upon
1:36:46 this thing called the worldwide
1:36:48 web and I ended up co-founding
1:36:51 agency.com which was one of the first
1:36:53 digital agencies and in the
1:36:55 mid90s and so I got to be there for one
1:36:58 of these revolutions and it was [ __ ]
1:37:03 magical and I and I I I seriously
1:37:06 thought that would be the last time in
1:37:07 my life right and then Along Comes Ai
1:37:10 and I think this one I think I I think
1:37:14 AI quite
1:37:15 frankly pales in comparison to
1:37:19 most human advancements or or or it it
1:37:23 makes most human advant advancements
1:37:26 pale in comparison to what's coming um
1:37:29 we are so lucky to be alive right now
1:37:32 when this thing is is is literally being
1:37:34 born to to to put it in
1:37:39 perspective take in that anyone that is
1:37:42 in any
1:37:43 way curious about AI right
1:37:47 now we're going to be the only ones on
1:37:50 the planet for some time that understand
1:37:54 what life was like before Ai and
1:37:57 after that's a remarkable thing to be
1:38:00 able to say in my day you actually had
1:38:02 to use your brain to think that's weird
1:38:05 Grandpa what was that like well it was
1:38:08 pretty
1:38:14 exhausting oh
1:38:19 man I'm shocked at the amount of people
1:38:22 that still won't talk about it I know li
1:38:24 listen that's why this channel exists so
1:38:26 do me a favor if if you like this
1:38:28 channel if you like the style I you know
1:38:30 listen I'm not I'm not everyone's cup of
1:38:32 tea but if you enjoy this please let
1:38:34 people know about this the whole purpose
1:38:36 of I go live five nights a week for
1:38:39 whatever it is two two two hours a
1:38:43 night because I think it's so important
1:38:46 that that people actually get their
1:38:49 hands on this [ __ ] and play with it play
1:38:52 with it you don't have to learn it you
1:38:53 don't have to take it seriously just
1:38:54 play with it just understand what it
1:38:56 makes [ __ ]
1:39:00 possible because we're
1:39:03 lucky that we get to play with this [ __ ]
1:39:06 we're the beneficiaries of like 8090
1:39:10 years of people trying to get to the
1:39:12 point that these things can do what they
1:39:15 do and and it was just sort of plopped
1:39:18 in our lap chat GPT on November 30th
1:39:22 2022 was like here
1:39:25 all of you that are not scientists have
1:39:27 had
1:39:28 it it's [ __ ] remarkable anyway I'm
1:39:33 sitting here in 2028 with my AI embedded
1:39:35 neural chip reminiscing about how
1:39:37 adorable it is that people are nostalgic
1:39:39 for the days when they when they have a
1:39:41 question first before finding the answer
1:39:44 exactly remember when we were not used
1:39:47 allowed to use calculators that's we're
1:39:49 we're in the same thing although these
1:39:51 are pretty fancy [ __ ] calculators all
1:39:53 right I'm going to get out of here I
1:39:54 hope you had fun tonight I know it so
1:39:57 one of the nicknames of the channel is
1:39:59 chat add and I hope I lived up to that
1:40:02 tonight I I live up to that most
1:40:06 nights so so peace out keep coming back
1:40:10 uh come to the um AI Salon meet and
1:40:13 greet tomorrow night it's from 5: to 7
1:40:16 PM mountain time and then I'll go live
1:40:19 after that as well so I might be a
1:40:20 little late tomorrow night but it's
1:40:22 generally 8:00 p.m. mountain time is
1:40:23 when I go live here all right so peace
1:40:26 out everybody and we will uh see you
1:40:28 tomorrow